Reversible Information Hiding In Encryption Domain Based On Multi-msb Embedding And Extreme Value Prediction

With the rapid development of cloud computing and other applications,more and more data are stored in cloud servers in the form of encryption.As an important branch of information hiding technology,reversible information hiding in encryption domain is a technology that can perform information embedding in encrypted multimedia data,which has the advantages of both traditional encryption and reversible information hiding.Reversible data hiding in encrypted image can not only protect users’ privacy,but also achieve the purpose of image marking or authentication.This paper studies reversible information hiding in encrypted image from the perspective of increasing embedding capacity.The main work and innovations are as follows:(1)A reversible information hiding algorithm based on additive homomorphic encryption and multi-MSB embedding is proposed.Firstly,the original image is divided into non-overlapping pixel blocks and each pixel sub-block uses the same key for additive homomorphic encryption,so that the correlation of pixels in the sub-block can be preserved.To improve security,the encrypted image is Arnold transformed in blocks.Then,by analyzing the difference between the pixel in the block and the predicted value,whether the sub-block can embed information and the capacity of the embedded information are determined.Finally,in embeddable sub-blocks,A small amount of LSBs(Least Significant Bit)save prediction difference ensures the image content is completely reversible,and the redundant multi-MSB embed the secret information in the way of bit replacement.In the decryption stage,the secret information can be accurately extracted from multiple MSB positions of pixels in the block,the image content can also be recovered losslessly according to embedded MSB numbers,predicted values and predicted differences.The proposed method can obtain a very large embedding capacity.Experimental results show that the average embedding rate of BOWS-2 database can reach 2.58 bpp(bits per pixel),which is better than other algorithms.(2)A high capacity reversible information hiding algorithm based on extreme value prediction is designed.First,the content owner divides the original image into non-overlapping sub-blocks of the same size and uses sub-block’s minimum pixel value(Min)as the predicted value of other pixel values in the sub-block.The difference between the maximum and minimum values is converted to a binary representation,and the length of the binary representation is taken as the encoding length of the difference between other pixels and Min in the sub-block.Secondly,the rearrangement of Min and difference is the compressed image content.When the encoding length of difference value is small,this method can vacate larger embedding space.Finally,the compressed image content is encrypted by stream encryption.The reserved space is located at the end of the compressed image,so the data hider can conveniently store secret information in the reserved space.When the receiver receives the marked encrypted image,the secret information can be extracted and the original image can be restored according to different keys.Experimental results show that the proposed algorithm outperforms many existing algorithms in embedding capacity performance.